Critical Incident Technique (CIT) is one of the oldest task analysis techniques and can be used in two ways. The first is to identify the critical components of a job or task (including exceptionally good performance as well as ordinary or ‘nominal’ performance). This means identifying (often through observation and interview) what needs to be done, and what should not be done. The second purpose more often used in safety circles, and the focus of this Fact Sheet, is to examine how people performed during challenging circumstances, e.g. during an incident or abnormal event. This can include: what happened; whether anything went wrong; how was it recovered; and what factors made it difficult or helped it succeed.
CIT can be further elaborated to ask those involved in the incident as to what could have made it better/worse, thus exploring human performance in variations of what actually happened. However, mostly CIT focuses on the facts as remembered, corroborated by other factual details wherever possible.
A weakness of the technique is that it is subjective, and ‘everyone is a hero in their own story.’ Nevertheless, often personal renditions of events can be compared to known facts and also compared to other people’s remembrances (i.e. others who were involved in the event), so the ‘hero’ biasing effect can be countered.
CIT can be done as a one-to-one interview, or one or two people interviewing a group who were involved in the incident. An advantage of group interviews is that other details may come to light during the discussion, as no one person has perfect recall. Group interviews can also reduce the hero bias. The disadvantage of group interviews is ‘group-think’ wherein everyone gravitates towards a unified version of what happened, usually a version that protects the team involved and is more conservative than what actually happened. The obvious solution (see illustrative example) is to conduct some single-person interviews first, and then a group interview.
(Flanagan, John C. Psychological Bulletin, Vol. 51, No. 4, July 1954.)
Kirwan, B. and Ainsworth, L. (1992) A guide to task analysis. London, Taylor and Francis
CIT is a flexible interview approach – as such it requires no prior training.
A set of questions are asked. These are best undertaken with open-ended questions, and the interviewees should be encouraged to state what happened and why in their own words (their narrative).
CIT can be single or group interview, or both.
A set of ‘factors’ may be used to explore what was affecting human performance during the event. These factors can be used to inform Human Reliability Analysis / Risk Analysis if that is to be carried out.
The answers require no special codification, though they often feed into other task analysis methods such as Hierarchical or Tabular Task Analysis, or Operations Sequence Diagram (see aviation illustrative example), or directly into designer consideration.
CIT can really add another dimension to a task analysis that is purely based on what people ‘should’ do. As such it adds important ‘colour’ and understanding of how stress plays a part. It can highlight what information (e.g. from displays or communications or observations) the people in the incident really paid attention to, and what they ignored, which is valuable information for a designer or procedure writer. It can also give useful insights into how actions between different people in the incident were coordinated, and the time it took for things to happen or be affected. It may identify issues or factors the designers, procedure writers and/or trainers had not previously considered. Incidents and accidents are where the design, procedures and training – in fact all the Human Factors in the system – are tested ‘through fire’, and you see what helped them make it through such challenging events.
More generally, CIT can also prevent ‘hindsight bias’, whereby people (after the event) wonder why those involved in an incident or accident didn’t simply ‘do the right thing.’ This is usually because in the heat of the moment, the ‘right thing’ was not clear. A relevant example of this was the landing of a plane on the Hudson River after a bird-strike to both engines. The flight crew were initially criticised for not trying to fly to the nearest airport. Only after further simulations, taking into account the required thinking time to make a decision, was it realised that if they had done so the plane would have crashed into a populated area with considerable loss of lives. A CIT would have identified this thinking-time component, often underestimated by designers and procedure-writers.
The disadvantage of CIT is that it is subjective
. Therefore, other more objective methods may be employed (e.g. prototyping and simulation) to verify any design decisions informed by CIT. Similarly, CIT focuses on one or two events that have actually happened, and insights should not be give disproportionate weight, as there are likely to be many other potential incident /accident scenarios that have yet to happen. Nevertheless, the fact that something has actually happened means it could happen again, and so some kind of design or other restorative action will be required. CIT will help ensure that the restorative ‘fix’ addresses the human element, making it more acceptable to future operators of the system.
How It Works
The basic approach is very simple, and involves asking a question along the following lines:
If it relates to an incident or accident that has actually happened, the opening is along the lines of the following:
Please tell us in your own words what happened.
Typical follow-up questions are as follows:
- Can you tell me what you were doing before the event? Were things quiet? Busy?
- When and how did you notice that something was happening?
- How did you react?
- How did others around you act?
- How was the situation resolved?
- How did it affect you at the time, and how did you feel about it afterwards?
- Were you surprised or startled?
- How unsafe did it feel, on a scale of 1 to 10?
Once the facts have been established, it is possible to ask about the factors that may have influenced their performance. This is best done first with an open question, so that they tell you in their own words, and then they can be shown a list and asked if any of the following factors listed affected them.
- Unfamiliarity with handling a wake event
- Workload (high or low – please specify)
- Team coordination
- Vigilance / fatigue
- Visual circumstances
- Weather conditions
- Time of day / night-time
The final questions shift the interview from what did happen to what could have happened:
- What could have made it worse on the day?
- What might have made it better?
- If a similar event happened again, would you do anything differently?
- What would you have liked to have been aware of:
- before the event
- during the event
- after the event
- Anything else you’d like to add?
CIT was used to interview three pilots who had experienced Wake Vortex events, wherein their aircraft encountered the wake from an aeroplane ahead of them, creating moderate to severe turbulence for their aircraft. Specific variants of the CIT generic questions were asked:
- Were you aware of the other aircraft?
- Did ATC warn you?
- Could you see their condensation trail? Could you see the other aircraft?
- Did you consider you might encounter a wake or was it a sudden surprise?
- Did the autopilot disengage automatically?
- Did you take manual control? How easy was it to control the aircraft?
- Did ATC give instructions/information during or after the event?
- Did you inform ATC?
An example of the type of material gained from one of the interviews is as follows:
The event occurred during daytime, and during a period of low alertness.
The captain saw the contrail of the aircraft ahead “coming down” which was a “surprise”.
The captain told the First Officer (FO) to put the seat belt sign on and move the seat forward to get ready for the encounter. No cabin crew warning was given.
The rolls of the vortex were visible: “rotating tubes” in clear air.
On a scale from 1 to 10 (10 being the most unsafe), the captain categorized the event as a 7 or an 8, mentioning “significant stress”.
The roll was “sensed” as a 15°-degree angle of bank, although it was 8°.
The startle of the encounter had passengers screaming but no injuries were recorded.
After the event, the flight crew contacted ATC.
Currently the pilots build an image of the surrounding airspace by using R/T information, contrail, TCAS information and wind information + geometry of aircraft pairs, if they are not informed by ATC.
By having prior ATC instructions/warnings, e.g. a verbal alert several minutes before the possible encounter, pilots believe they would overcome or better prepare for encounter as it may occur during “low alertness” states.
Training should expose flight crew to en-route wake encounters both theoretically and practically. The pilots think the training will most probably focus on recovery as it is hard to simulate the “startle” effect.
The top 6 factors (no priority order) identified by the group of pilots were as follows:
- Seatbelt sign on / cabin crew info (would minimize the outcome of the wake in terms of passenger and cabin crew disturbance /potential injuries)
- Covering controls (being ready to take control)
- Vigilance / Situational Awareness (SA) / circadian rhythm and arousal (monitoring throughout the flight depending on flight crew alertness)
- Startle effect (slower reactions or over-corrective input)
- Robust Autopilot (depending on aircraft type it might disengage faster)
- IMC (in cloud) / night (less “outside” view / lower alertness)
The CIT in this case was part of a sequence of task analyses for the Wake Vortex study. You can see how it informed the Operational Sequence Diagram (OSD) by reviewing the case study in the Factsheet for OSDs.
While approaching a harbour, a bulk carrier was met by an escort tug who was to assist the vessel to a mooring station. When approaching the harbour entrance, the vessel experienced a sudden loss of steering. A loss of steering is an unexpected event that can be manifested by several mechanical issues (e.g. blackout, engine gear box failure, steering engine failure on rudder, rudder jam, loss of starting air pressure (due to many consecutive manoeuvres).
Specific variants of the CIT generic questions were asked:
- Were you aware of any ongoing repair or maintenance orders that might affect steering control?
- Were you in communication with the engine control room (ECR)?
- Were you aware of any bridge alarms that would indicate abnormal conditions?
- Were you confident about the manoeuvring safety zones and traffic conflicts nearby?
- Did you communicate the issue at first indication to others on the bridge or in the ECR?
An example of the type of materials gained from crew interviews identified reasons for steering failure and how human factors, automation and work practices may provide procedural, pedagogical and/or mediating actions.
The emergency generation should start and provide enough power for essential equipment like the steering engine. If the power management system is working and there are auxiliary engines ready and in automated mode, they should start, connect and equipment should by sequence automatically start. The emergency generator stops automatically when auxiliary engines are up and running and can take the load.
Routines and check lists should be in place to get everything up and running. Communication with the bridge is an immediate action to inform about the situation.
If it is a critical situation, narrow passage, arrival or departure the ECR must be manned by the chief engineer who usually is responsible for informing the bridge about the progress.
Problems with the steering engine:
A hose could break causing a rapid drop in hydraulic pressure. Usually there are two electrohydraulic motors and in critical situations two are running. If one stops the other should automatically start.
Bridge should be informed about this incident.
Loss of remote steering:
If the remote steering from the bridge is not working the steering engine can be operated from the steering gear room.
The hydraulic valves are then hand operated according to the bridge instructions. A rudder indicator is fitted in the steering gear room and the engineer is following the instructions from the bridge.
The communication is through a fixed mounted emergency phone or using VHF.
Checklists and routines need to be practiced regularly (re. SOLAS)
Problems with the rudder:
Mechanical failure causes the rudder to stick.
Problems with the main engine:
Numerous reasons for the main engine to stop.
- Fuel quality
- Shutdown alarm on an auxiliary engine
- Could also be a mechanical failure which forces a stop of engine
- Leaks of fuel, lubricating oil, cooling water etc.
- Problems with the gear
- Mechanical failure
- Hight temp lubricating oil, high temperature lubricating oil.
Problems with basically any critical transmitter
The top factors (no priority order) identified by the bridge and ECR teams:
- An emergency plan with the escort tugs (if within arrival and departure protocols) should be reviewed (similar to a pilotage plan)
- Better communication between bridge and engine control room
- Clear understanding of operations and detection of abnormal sensor data outputs
- Training for immediate control (emergency steering system of the steering engine)
Bridge team better aware of the sensor mimics and alarms from the Engine Dept.